Marine, automotive, trucking, rail, aerospace, defense, recreation, chemical, infrastructure, and other industries look to composite materials to take advantage of their unique properties, especially being corrosion-free or corrosion-resistant and having a high strength-to-weight ratio. Composites are also resistant to fatigue and chemical attack. They offer high strength and stiffness potential in lightweight components. There is a need, however, to develop composite manufacturing processes that dramatically reduces the cost of composites, especially large structures, while retaining their high strength and stiffness.
Resin-impregnated fibrous materials (prepregs) generally are placed on a forming mandrel (“laid up”) by hand or machine using tape, fiber tows, or cloth. Composites also have been made using filament winding. Debulking is required between plies in a laminate to remove air before the layups are vacuum bagged (i.e., enclosed in an inert atmosphere under vacuum to withdraw emitted volatiles released during cure of the resin) and consolidated in autoclaves or presses to achieve high fiber volume components. The prepreg materials typically are expensive (especially those using high modulus carbon fiber). The raw prepreg materials have limited shelf lives, because the resins react at slow rates (“advance”) at ambient temperature. Advance of the resin adversely effects the properties of the resulting composite. Working with prepreg also often results in considerable material waste.
The autoclaves and presses used for consolidation are expensive capital items that further increase the final, manufactured cost of the composite. Processing has to be centralized and performed in batches where the autoclave or press is installed. Loading and unloading the autoclave (a high temperature, pressurized oven) usually become the rate limiting steps. The location of the autoclave dictates where the composites will be made, so the flexibility of the process is impaired. A dedicated workforce and facility are required, centered around the autoclave.
As mentioned, prepregs have a limited shelf life. In some formulations, the resin is carried onto the fiber as a lacquer or varnish containing the monomer reactants that will produce the desired polymer in the composite (i.e., prepregs of the PMR-type). In other formulations, the resin is a relatively low molecular weight polymer that crosslinks during cure to form the desired polymer. The resin is held and used in its incomplete state so that it remains a liquid, and can be impregnated onto the fiber or fabric. Reaction of the monomer reactants or crosslinking of the polymer (i.e., its advancing) prior to the intended cure cycle adversely impacts the quality of the composite.
Liquid molding techniques such as transfer molding, resin film infusion, resin transfer molding, and structural reaction injection molding (SRIM) typically require expensive matched metal dies and high tonnage presses or autoclaves. Parts produced with these processes are generally limited in size and geometry. The conventional liquid molding resins do not provide the necessary properties for many applications of composites.
Open mold wet layup processing can make large composites using a liquid molding process with minimal capital equipment, single sided tooling, and often can use lower cost materials than prepreg. The quality and uniformity of the product, however, varies considerably and the best composites are still relatively low quality. The process also tends to be unfriendly and presents hazards to workers because of their risk of exposure to the solvents and resins.
Our double bag vacuum infusion (DBVI) process solves a number of problems encountered with previously developed, nonautoclave, single bag, liquid molding techniques, such as those processes described in U.S. Pat. Nos. 4,902,215 (Seemann) and 4,942,013 (Palmer). In the Seemann single bag technique, preferential flow and pressure is induced in the flow media above the fiber preform. The driving force is a pressure differential or head pressure created primarily by drawing down the pressure inside the bag using a vacuum pump. Atmospheric pressure on the resin feed pushes resin into the bag through an inlet tube. Resin entering the bag encounters the flow media used to channel the resin to the underlying fiber preform. Resin flows laterally through the flow media over the preform and, subsequently, downwardly into the preform. The preform has the lowest permeability to flow (i.e., the highest resistance to the flow of resin).
Once the liquid media (i.e., resin) is pulled (i.e., flows) into the preform, we have observed that the single bag tends to relax behind the wave front (i.e., the foremost portion of the resin that is moving into the preform within the bag). When the flow media is full or partially full of resin, we believe that the bag slowly relaxes and moves away from the flow media presumably because the flow path of least resistance becomes a path over the flow media between the flow media and the overlying bag. Relaxation increases the enclosed volume around the preform, which becomes filled with resin. The farther away from the leading edge of the wave front, the more the bag tends to relax. We have observed that the composite in areas where the bag has relaxed can have lower fiber volume, poor fiber volume control, and lower mechanical properties than desired, because excess resin has filled the enlarged volume. The bag relaxation can produce a change in the intended thickness of the composite, so that in localized areas where relaxation has occurred the composite is thicker than intended.
In the ensuing discussion, we will compare the Seemann and Palmer processes with our preferred double bag process of the present invention.
Our preferred double bag vacuum infusion process circumvents the Seemann (single bag) problems in that the inner and outer vacuum bags independently control the resin feed. The double bag provides a caul effect. The bleeder and breather sections are completely isolated. With this approach, the bag is never able to relax behind the wave front and the resulting composites have higher fiber volumes on average (with more precise control) and have uniform thickness with constant thickness preforms. Our process eliminates the bag relaxation defects we observed with the Seemann process.
During relaxation, we observe that resin pools inside the bag. Pressing on the pool, we feel a soft, spongy, loose area different from the feel where relaxation is not occurring. The bag stretches and the volume under the bag increases. In circumstances of relaxation, we have observed that pressurizing the resin feed above atmospheric increases the relaxation, so the phenomenon appears to be tied to the pressure differential and the driving force for resin flow, as we would expect. Adding a second vacuum bag (separated from the first bag with a breather) makes it harder for the “double bag” to relax. Therefore, we can use a higher differential pressure to move the resin than might best be employed with a single bag. The “double bag” becomes a means to reduce flow over the filled flow media because the vacuum bag effectively is thicker. The “double bag” also provides increased vacuum integrity because it provides a redundant, second bag to counter any leaks in the first bag.
In Boeing's “Controlled Atmosphere Pressure Resin Infusion” (CAPRI) process, Jack Woods et al. control the differential pressure by reducing the pressure below atmospheric in the resin feed tank. In the CAPRI process, a vacuum pump evacuates the volume under the vacuum bag while, simultaneously, reducing the pressure over the feed resin. Pressure in the vacuum bag might be .sup.31 20 inches Hg below atmospheric and .sup.-5 inches Hg in the feed pot for a differential pressure to drive resin infusion of 15 inches Hg (.about.0.5 atm).
The Palmer process attempted to isolate the bleeder and the breather sections by placing an impervious film between the flow media and the breather inside the single bag. Unfortunately, this technique did not allow complete isolation. Once the liquid medium reached the vacuum end of the assembly, the flow media and the breather were connected. As a result, the resin began to wet the breather and to flow back toward the resin source over the membrane because this path had higher permeability than flow downwardly through the preform.
Our preferred double bag process allows fiber volume percentage or fraction in the composite to be increased 5-10% higher than we have been able to achieve with the single bag technologies of Seemann and Palmer. An increased fiber volume is critical to achieve an aerospace grade composite that has properties competitive with conventional vacuum bag/autoclave prepreg technologies commonly used in aerospace. Aerospace composites have superior ‘specific strengths’ which are achieved by optimizing (making as high as possible) the fiber volume fraction. Aerospace composites have superior ‘specific strengths’ which are achieved by optimizing (making as high as possible) the fiber volume fraction. Our process achieves a targeted fiber volume within a close tolerance of acceptable fiber volumes by regulating the vacuum of the inner and outer bags during infusion. Using end game thermal infusion strategies, our process improves preform nesting, fluid drawdown, thermal vacuum debulking and real time mass balance control. Our process has extremely high vacuum integrity.
In any vacuum impregnation process, vacuum integrity is essential to produce high quality composites consistently. Leaks in the bagging seals, resin ports, or vacuum ports will permit air to enter into the bag. Air causes the preforms to swell and reduces the fiber volume fraction by increasing the spacing between fibers. Composites made with leaking bags will typically have one of more of the following problems: high void content, surface porosity, low fiber volumes, or excessive thickness. Parts often need to be scrapped; they cannot be repaired.
In vacuum bag processing, one side of the structure is tooled and the other is defined, at least in part, by the bagging materials used over the layup. Bag side roughness and mark off is a common problem experienced with prepreg processing and bag liquid molding processes. Cauls and intensifiers are often used on the bag side of the laminate to improve surface finish. These surface enhancements, however, are not particularly effective in the Palmer or Seemann process because of the flow medias used. The coarse knotty knit flow media and the bag offset materials described in the Seemann process result in bag side mark off on the parts even in the presence of peel ply separator. Mark off occurs because of localized high pressure at the knit knots or bag offsets with relatively low pressure in surrounding areas. The uneven pressure distribution produces a relatively lumpy bag side surface. Fiber volume and fiber content varies.
Palmer uses glass bleeder cloths to form part of his flow media pack. Layers of dry glass cloth tend to bunch, buckle, and bridge under vacuum causing severe mark off problems even on simple geometric part configurations, not to mention the complications that arise in more complex assemblies.
Used in the Seemann process to achieve rapid lateral flow, thick flow media and bag offsets create relatively large volumes that will ultimately fill with waste resin. In Palmer's process, the flow media, the thick glass packs, and also the glass bleeder diapers waste resin. Palmer also loses resin when it flows beyond the end of the infusion and wets into the breather, as we discussed. We seek to minimize resin waste.
In our preferred process, resin losses in the flow media are reduced because of its low profile and relatively small open volume. Our process also allows for simple purged resin reclamation and recycling without the risk of bag relaxation or the need for continuous resin purging with fresh resin to infuse difficult preforms. Our preferred process conserves resin and reduces cost measurably when working with expensive resin systems, as is common for aerospace applications.
Neither Seemann nor Palmer describes how to produce complex assemblies such as contoured skins with blade stiffeners, where the plumbing requirements are complex. Each stiffener requires an active vacuum line attached at the top of the stiffener to draw the resin up into the stiffener. When there are a large number of stiffeners, the plumbing quickly gets complicated. Each connection requires flawless seams with the bag to preserve vacuum integrity. In our process, some stiffeners can be effectively infused without using active vacuum lines. Inclined infusions where the resin is introduced at the lowest point and pulled up the preform to the highest point can effectively wet out stiffeners running in the flow direction and in some cases other directions as demonstrated in our TYCORE™ sandwich panels.
Our process also can install passive vacuum chambers (PVC) inside the inner bag. Perforated tubes, spiral cut tubes, springs or other open containers are placed above stiffeners or other areas where flow is desired (E, FIG. 8 or 9). The resin or liquid is pulled into these chambers until they fill. The PVCs also provide some purging capability for removing air from preforms.
By “wet out” we mean infusion of the desired amount of resin into the preform to achieve the desired fiber volume in the composite.
The Seemann and Palmer processes can produce parts of almost unlimited length but are limited with respect to part width. Seemann's process can generally produce wide simple shells because Seemann uses flow media having high permeability and bag offsets. Palmer's process is somewhat more limited because it relies on an edge feed method and uses flow media of lower permeability. At some width, however, both the Seemann and Palmer processes require additional feed lines to reduce resin drag and pressure drop in the system, especially where flow on a skin is interrupted with stiffeners. Stiffeners create choke points when the resin is flowing transverse or at an angle relative to the direction of the stiffener. Because of tooling constraints, dimensional control requirements, and shape discontinuities, care must be taken to place flow media materials in stiffener locations properly.
A variety of dry preforms are available for constructing infused components. Both Seamann and Palmer use dry preforms. The options include standard weaves, warp knit materials, 3D braids, 3D woven materials, stitched preforms, Z-pinned preforms, continuous strand mats, and chopped fiber preforms. Many dry preform materials are fragile, easily distorted, damaged, or frayed from simple customary manufacturing operations. Distinct ply dropoffs, part tailoring, and net shapes are difficult to achieve in complex finished parts made from dry preforms. Dry preforms also tend to have excessive bulk for layup of complex shapes where bulk must be minimized to eliminate wrinkling and bagging issues. To compound the problem, layers of the dry materials cannot be debulked and consolidated effectively because of their poor adhesion to other dry plies or to other materials. Offline detail preform fabrication is ineffective. These characteristics make dry preforms difficult, if not impossible, to use in many complex applications. Therefore, tackifier or binder technologies for treating dry preforms with resin necessarily become key elements of almost any liquid molding technology system. The binder must not restrict resin flow or preform consolidation, must be compatible with the infusion resin, and must not produce loss in strength. The process of applying binder or tackifier produces a preform similar to those used in conventional resin transfer molding.
In our preferred process, again, we have developed a unique spray impregnation process to apply the binder or tackifier to the dry fiber preform to produce high tack with low binder content. Desired binder content ranges from about 1 to 10 wt % (i.e., by weight) but typically are from about 3-7 wt %. The desired weight percent depends on the weight and thickness of the preform and the natural or inherent degree of tack in the binder.
Adding solvent to the semi-solid viscous resins is useful as binders produce solutions suitable for spraying. The solvents should have room temperature tack and be compatible with the infusion resin selected. For cyanate ester infusion resins, we typically use CIBA's M-20 semi-solid cyanate ester resin that is extremely tacky at room temperature. Some semi-solid resins with no room temperature tack can be used if they develop tack when heated, for example, 5250-4 RTM bismaleimide resin. The solutions sometimes require catalysts for resin activation. For more latent spray formulations, the catalysts can be eliminated or reduced from the mix to allow higher temperature vacuum debulk operations without adversely advancing the degree of cure of the binder. Binder contents can be increased at ply edges to provide greater dimensional integrity and less edge fraying. The binder might also incorporate thermoplastic or rubber toughening agents for improved damage tolerance and ballistic survivability.
The preferred binder formulations typically have high or very high resin solid contents of 80% by weight or more. The solvent or carrier can be MEK, MIBK, other organic solvent capable of dissolving the semi-solid resin, or, possibly, water. Solvent volatility can be altered and used to control or to adjust tackiness and to change drying time. High solids content, high spray viscosities, and dry film spray parameters are used in conjunction to form uniformly distributed small resin spots that rest on the exposed surface of the preform. The preferred spray parameters minimize solvent emissions, increase transfer efficiencies, allow automation, and maintain maximum preform tack with the least amount of deposited resin and loss in preform permeability.
With these tackified preforms we have demonstrated the ability to produce complex structures such as intersecting blade stiffeners, Pi joint stiffeners, and complex contoured skins with curved blade stiffeners. The binder technology makes it possible to net mold certain features such as blade stiffeners. Vacuum bag, room temperature debulking can produce soft, pliable, tackified preforms. Heated vacuum debulking can produce semi-rigid preforms suitable for precision trimming to close tolerance.
With the Seemann and Palmer processes, the resin must be gelled immediately after part infusion. If you leave the vacuum active on the part, resin from the source is pulled through the preform during the gelling. The resin supply must remain connected to prevent the part from being depleted in resin, For most resins, gelation is initiated thermally. Heating the part to gel the resin in the preform also heats the bulk resin which can lead to a hazardous exothermic condition, including evolution of toxic smoke.
If you close the vacuum and feed lines for the bulk resin prior to heating the preform, leaks might cause air to bleed into the bag. This bleed often produces defective parts that have high void content. The part may swell to create low fiber volume components or, more typically, ones having voids or porosity. The Palmer process requires almost instantaneous gelation, but excessively rapid gelation often produces brittle resin matrices. Many common resins, such as low temperature cure epoxies for high temperature applications, cannot be gelled rapidly.